The '3904 is the NPN (2n2222) and the '3906 is the PNP (2n2907). You can run it off of a cap with the appropriate current limiting resistors, but the relationship between how long a cap with 10 times the farads of a smaller one will last isnt linear, for several reasons.
Hi,
SURE! The discharge time is not linear, its exponential... Use this formula:
Vf = Vo(e^-t/RC)
where Vf = final voltage
Vo = initial voltage
t = time in seconds
R = resistance in ohms
C = capacitance in FARADS (not microF)
e = "natural log base" = approx. 2.718
Hi,
SURE! The discharge time is not linear, its exponential... Use this formula:
Vf = Vo(e^-t/RC)
where Vf = final voltage
Vo = initial voltage
t = time in seconds
R = resistance in ohms
C = capacitance in FARADS (not microF)
e = "natural log base" = approx. 2.718
If you go back and read the whole thread (a daunting task, admittedly), you will find that he wants to power the circuit from a big cap, and is wondering how long the circuit will continue to flash as the big cap discharges through the circuit.
If we use 2.2k ohms as the resistance of the circuit (you have to discharge the monster cap through a limiting resistor), and charge the 1 Farad cap to 12 volts, and stop the clock at 5 volts, I get about 1926 seconds, or just over half an hour.
Use the formula I mentioned before, solving for t using natural logs:
I think you did well, thanks. I don't think I will end up having a useful purpose to run this off a cap, but its fun to play with anyway. I appreciate all the help for such a meaningless project. Although, I have learned a lot, so I guess its not wasted.
If we use 2.2k ohms as the resistance of the circuit (you have to discharge the monster cap through a limiting resistor), and charge the 1 Farad cap to 12 volts, and stop the clock at 5 volts, I get about 1926 seconds, or just over half an hour.
Use the formula I mentioned before, solving for t using natural logs:
Joecool, it gets better.
Hyedenny, where did you come up with 2.2k, 12V, and 5V? Last I heard, he wants to use a 1 Farad cap in place of a battery as the POWER SUPPLY for the flasher, starting with a 5V charge.
I think this would be hard to come up with analytically. I simulated it on SwitcherCAD, and it oscillated for 1000 minutes. It might not be useful for that length of time, because, while the period (2.2 sec) and pulse width (10ms-20ms) remained relatively constant, the LED current, as you might conclude, did not. The peak current was only about 3.5ma just before it stopped oscillating.
Well, it was a nice lesson, but I suspect Joe thought it was the answer to his question, when in reality it had very little relevance. The flasher is a nonlinear load. While running the sim, I fiddled around with coming up with an equivalent load. A 31k resistor in series with an LED matched the discharge curve of the flasher circuit almost exactly.